Transcript
DUMB AND DUMBER IN MACROECONOMICSRobert M. SolowSo how did macroeconomics arrive at its current state? The answer might provide a lead as to where it ought to go.The srcinal impulse to look for better or more explicit micro foundations was probably reasonable. It overlooked the fact that macroeconomics as practiced byKeynes and Pigou was full of informal microfoundations. (I mention Pigou todisabuse everyone of the notion that this is some specifically Keynesian thing.)Generalizations about aggregative consumption-saving patterns, investment patterns,money-holding patterns were always rationalized by plausible statements aboutindividual--and, to some extent, market--behavior. But some formalization of theconnection was a good idea. What emerged was not a good idea. The preferred modelhas a single representative consumer optimizing over infinite time with perfectforesight or rational expectations, in an environment that realizes the resulting plansmore or less flawlessly through perfectly competitive forward-looking markets for goods and labor, and perfectly flexible prices and wages.How could anyone expect a sensible short-to-medium-run macroeconomics tocome out of that set-up? My impression is that this approach (which seems now to bethe mainstream, and certainly dominates the journals, if not the workaday world of macroeconomics) has had no empirical success; but that is not the point here. I startfrom the presumption that we want macroeconomics to account for the occasionalaggregative pathologies that beset modern capitalist economies, like recessions,intervals of stagnation, inflation, stagflation, not to mention negative pathologieslike unusually good times. A model that rules out pathologies by definition is unlikelyto help. It is always possible to claim that those pathologies are delusions, and theeconomy is merely adjusting optimally to some exogenous shock. But why shouldreasonable people accept this? During the past three years, unemployment hasincreased by three million with real wages stagnant and productivity growing, possibly abnormally fast. Capacity utilization has fallen by 10 percent, with trivialinflation and some prices falling. Real business investment in equipment peaked inthe third quarter of 2000, fell by 20 percent to the first quarter of 2002, and has risen by a scant five percent since then. Is this a stagnation pattern? Does it reflect large-scale, perhaps irrational, overinvestment in the 1990s? Should it not be studied assuch? Why should anyone take it as the solution of an Euler equation? It would not be hard to imagine a better path for the economy. Why should the burden of proof fallon those who see an ordinary standard pathology here? The odd thing is to regard thishistory as the working out of an other-worldly model.What is needed for a better macroeconomics? My crude caricature of theRamsey-based model suggests some of the gross implausibilities that need to beeliminated. The clearest candidate is the representative agent. Heterogeneity is theessence of a modern economy. In real life we worry about the relations betweenmanagers and shareowners, between banks and their borrowers, between workers andemployers, between venture capitalists and entrepreneurs, you name it. We worryabout those interfaces because they can and do go wrong, with likely macroeconomicconsequences. We know for a fact that heterogeneous agents have different andsometimes conflicting goals, different information, different capacities to process it,different expectations, different beliefs about how the economy works.Representative-agent models exclude all this landscape, though it needs to beabstracted and included in macro-models. 2I also doubt that universal rational expectations provide a useful framework for macroeconomics. One understands the appeal. Think of it this way: Herb Simon wassurely right about bounded rationality; no one would deny that most economic agentsare actually like that, and natural selection does not work fast enough to eliminatethem. Why did the notion of satisficing never catch on? I think it is because theassumption of complete rationality tells the modeller what to do, whereas boundedrationality only tells the modeller what not to do. That is not helpful. Somethingsimilar is true about rational expectations. If there were a nice parametric family of alternative ways to model expectations, it might catch on. Most of us would happilygo along with the notion of expectational equilibrium: if specific underlyingexpectations generate an outcome in which those expectations are systematically andnon-trivially violated, that situation can not be an equilibrium. It is what happens thenthat needs thought. The situations that agents need to anticipate need not even be probabilistic, surely not stationary. The popular device used to be adaptiveexpectations; that may have been inadequate. Maybe this is a case for the applicationof psychological research (and sociological research as well, because the formationof expectations is a social process). Maybe experiments can be designed.Heterogeneity across agents and classes of agents is certainly important preciselyhere. One would like a simple, definite way to proceed, if that is possible. A goodexample of the sort of thing I mean is the way the Dixit-Stiglitz model mademonopolistic competition easy. (The trouble is that we are dealing with anunobservable.)Although I am going to take this back in a moment, it is certainly worthwhilementioning the problems connected with real and/or nominal wage and priceinflexibility and its sources in market structure, limitations of information, humannature, the specialness of zero, etc. This is an old issue in economics, macro andmicro, and a lot of progress has been made in measuring and understanding it. Meresluggishness is part of the picture, and that is easily modelled, but there is surely morethat is less easily modelled. The devil finds work for idle hands to do, as you mayhave noticed. Now here is a peculiar thing. When I was in advanced middle age, I suddenlywoke up to the fact that my colleagues in macroeconomics, the ones I most admired,thought that the fundamental problem of macro theory was to understand hownominal events could have real consequences. This is just a way of stating some puzzle or puzzles about the sources for sticky wages and prices. This struck me as peculiar in two ways.First of all, when I was even younger, nobody thought this was a puzzle. Youonly had to look around you to stumble on a hundred different reasons why various prices and factor prices should be much less than perfectly flexible. I once wrote,archly I admit, that the world has its reasons for not being Walrasian. Of course Isoon realized that what macroeconomists wanted was a formal account of pricestickiness that would fit comfortably into rational, optimizing models. OK, that is aharmless enough activity, especially if it is not taken too seriously. But price andwage stickiness themselves are not a major intellectual puzzle unless you insist onmaking them one.The second peculiarity was that the path from nominal events to real 3consequences was not my idea of the fundamental problem of macro theory anyway.All along, I had been thinking--and this may be a Keynesian inheritance, though Idoubt it because I may have picked it up from Gottfried Haberler's Prosperity and Depression , where my generation learned about business-cycle theory before macroeconomics had been invented--that the main problem was to understand whyreal shocks that took the economy out of some satisfactory equilibrium led to such a prolonged and sometimes unsatisfactory adjustment. These are medium-run problems--the capital stock moves--and there clearly are medium-run fluctuations in modernindustrial economies. (This is documented for the U.S. in a recent paper by Cominand Gertler.)Keynes claimed to have found the way to account for this: he thought he hada theory of unemployment equilibrium. The reason adjustment took so long, or never really happened, is that the depressed state was actually an equilibrium. Most of ustoday think that Keynes failed in that effort; he lacked the tools. The exception wasthe case of wage rigidity, but we knew that all along. In my youth, we thought thatmacro-pathologies were disequilibrium phenomena, and then the puzzle was: why isthe process so slow?This choice between equilibrium and disequilibrium thinking may be a falsechoice. If I drop a ripe watermelon from this 15th-floor window, I suppose the whole process from t 0 to the mess on the sidewalk could be described as some sort of dynamic equilibrium. But that may not be the most fruitful--sorry--way to describethe falling-watermelon phenomenon.So I would hope that macro theory could get back to focussing on theadaptation-to-real-disturbance problem, without falling into the implausibilities of real-business-cycle theory. (Even RBC theorists may fight their way out of that paper bag.) The Ur-Problem may be: start in a situation of growth equilibrium ( not necessarily a steady state, but don't get me started on that one), and imagine a realshock, perhaps a failure of real effective demand (!). What happens next? That may be the story of the period from 2000 to now, the real shock having been massiveoverinvestment in response to unrealistic profit expectations (accompanied byaccounting swindles, just to make Joe happy).
